FM synthesis

anon45924993 · 2017-10-24T23:19:20.194Z

Yeah, it seems like it’s just a circuit design for a MIDI interface for a Pi. I don’t see any software there.

rick_monster · 2017-10-25T00:42:33.516Z

getting some pretty legit ‘vintage FM’ sounds now! (well I like them anyway)

so this is using the aleph as arpeggiator & a mostly bugfixed version of the fmsynth module. First minute is cycling through some presets I’ve programmed - then I start messing around with FM parameters pretty randomly leveraging aleph’s preset system to switch control dials between ‘banks’ of parameters.

Recording stops eventually after way too much twiddling… This is kind of exactly why i started this module - love my little 4-op yamaha to bits but the programming is a royal pain & requires cracking out my laptop, setting up the software etc, even then iirc, many of the ‘params’ don’t update realtime… so I just end up using the stock patches like everyone else!

Oh man if only we had usb hub support for avr32… I would love to get the MPC & aleph grid ops playing together!

capogreco · 2017-10-25T07:00:07.587Z

fyi: an elegant random FM instrument generator, coded in SuperCollider by James McCartney.

cmcavoy · 2017-10-25T19:31:31.401Z

Forgive this naive question - but is there a difference between linear / exponential FM between two analog oscillators and the smooth DX-7 FM that most of what we’re talking about here? It’s all just FM, right? Just different levels of control? Same principal as FM radio? Two frequencies modulating one another? Analog FM almost always sounds like noise, DX-7 sounds glorious. If we had an infinite amount of control over our analog gear, could we recreate DX-7 sounds?

capogreco · 2017-10-25T21:19:05.127Z

There is one big difference between linear and exponential FM, namely that linear FM retains the fundamental pitch of the carrier, and so is much more “musically” useful.

Mathematically, this is because with linear FM, the area under the curve of frequency vs time above the fundamental is equal to the area under the curve under the fundamental, so they sort of even each other out. With exponential FM, the area under the curve above the fundamental outweighs the area under the curve under it, which means as the modulation gets deeper, the perceived fundamental increases.

John Chowning was the pioneer of this technique, and wrote pretty extensively on it if you want to learn more

zebra · 2017-10-25T21:27:37.285Z

cmcavoy:

If we had an infinite amount of control over our analog gear, could we recreate DX-7 sounds?

if you have enough oscillators
if the modulation is through-zero
if the modulation targets phase and not frequency directly
if the modulation is linear
if the oscillators implement some method of phase sync
if the oscillators can be set to track each others pitch sufficiently accurately for your tastes

then yes.

i guess i just have different tastes because i find digital FM to be useful, but not exactly interesting. whereas analog FM is like some kind of window into the seething mysterious depths of the clockless universe. maybe i just like “noise”

anyway, dx-7, et al also has aliasing

widdly · 2017-10-25T21:49:03.625Z

There is an interesting patch in the Strange book where he uses a ring modulator as a multipler to compensate for the pitch shift in expo fm. I’ve had good results with it and I also discovered my homemade ring mod has inverted outputs.

cmcavoy · 2017-10-25T22:00:50.421Z

zebra:

then yes.

So what you’re saying is…yesssss. Got it, I’ll start working on it.

zebra:

analog FM is like some kind of window into the seething mysterious depths of the clockless universe. maybe i just like “noise”

I moved a bunch of my modules into a skiff to teach a ‘class’ but also to get my ten year old into modular. I’ve tried a few times in the past, but it never clicked for him and I think I dropped way too much theory on him. This time I backed off, walked him through a simple patch, then turned him loose. He did things I would have considered “mistakes” that ended up making noises i’ve never heard come from my modules before.

Last night we FM’d the STO from that skiff with a Makenoise DPO (both A & B oscillators), so that when he made a change, or I made a change it got realllll noisy. We added in modulation from cold mac, rene and wogglebug…it was super fun. This morning, when I dropped him off at school, I said “hey, last night we really created a whole lot of noise, it’s entirely possible we opened a rift in the universe and unleashed some sort of interdimensional being, so watch your back today.” So. Yes. “seething mysterious depths of the clockless universe” nailed it.

cmcavoy · 2017-10-25T22:01:44.458Z

capogreco:

John Chowning was the pioneer of this technique, and wrote pretty extensively on it if you want to learn more

Thanks, will check him out!

anon45924993 · 2017-10-25T22:02:23.944Z

cmcavoy:

He did things I would have considered “mistakes” that ended up making noises i’ve never heard come from my modules before.

That’s the beauty of embodied cognition. Think with your hands…

Then, think with your modules! (extended cognition…)

abalone · 2017-10-25T22:12:36.469Z

rick_monster:

I would love to get the MPC & aleph grid ops playing together!

probably off topic
here’s an art video we made back in the day
shown at usc fisher museum- laura’s art
https://news.usc.edu/2725/Aztec-to-Hi-Tech-A-New-USC-Art-Exhibit-Examines-the-Border-Culture-of-Bajalta/
tracks made with the mpc, and an goodwill mac with a scsi port that could run recycle into an s2000 akai sampler, and an orbit module (updated classic keys)… anyways

zebra · 2017-10-25T23:37:10.002Z

well if i understand it correctly, make noise DPO is more or less similar to the buchla 259, two of which (plus VCA matrix) perform the study linked above.

so they both are tricore with waveshaping, and have hard sync between carrier and mod oscs.

259 additionally has a sort of analog “soft sync” where instead of resetting the mod osc phase on the carrier sawtooth, it reverses its direction.

i think these sync mechanisms are important if you want very “stable” FM sounds.

and i think you can get these metallic timbres in analog that are relatively stable, but with that extra organic drift. and there is a wider harmonic palette because, really, it is easy to push dx- and tg- series into heavy aliasing, and its become part of “that sound.” bandlimited FM is just not an easy thing to do even on todays digital hardware; all you can do is constrain the modulation as a function of the input waveforms (heres a paper [edit: woop, nevermind, that’s a kind of different thing. what i had in mind was computing the bessel function and limiting mod depth, or bandlimiting mod signal, to constrain output of bessel func. this requires integrating the modulator so it’s not easy. deep rabbit hole.]).

then there is just they way that these systems actually converge on, or orbit around harmonic stability with the right settings, making glissandos and macro oscillations. in digital we can approximate with slews, delay, filtering, and added noise… but yeah, not the same.

i think if someone were to design an analog dual osc that was designed for “digital-like” fm synthesis, one thing that might help is a trigger input for sync. so at the start of a note you force the mod and carrier to be in sycnhronous phase, then let them go their merry way. is that a thing yet? if not, why not?

PS: speaking of chowning/stanford/yamaha, here’s a pretty fun piece my mom made, with max mathews performing. it uses tg77. she was really into fractal waveforms and made some kind of wild system of taking FFT of a fractal and turning it into tg77 sysex for additive waveforms. then using (approximately) irrational mod/carrier ratios related to the fractal dimensions and stuff. i think it’s quite musical but also not what we typically think of as “dx7 sound”

Discogs

Various - CDCM Computer Music Series Vol. 15: The Virtuoso In The Computer Age V

Label: Centaur Records - CRC 2190 • Series: CDCM Computer Music Series - Vol. 15,The Virtuoso In The Computer Age - V • Format: CD Compilation • Country: US • Genre: Electronic • Style: Experimental

the 90s

rick_monster · 2017-10-26T00:14:58.726Z

zebra:

bandlimited FM is just not an easy thing to do even on todays digital hardware

interesting point! so I wonder how different ‘DX7 FM’ sounds with something crazy like 16x oversampling…

sliderule · 2017-10-26T04:00:52.871Z

Not noted yet is Yamaha’s FM-X, available on the Montage. 8 ops and some spectral stuff I can’t remember plus knobs.

zebra · 2017-10-26T04:32:30.978Z

heavy oversampling is the typical way in softsynth-land, i think

kisielk · 2017-10-26T06:49:18.230Z

The Cylonix Shapeshifter uses an internal sample rate of 2 MHz (it’s implemented with an FPGA). FM is one of my favorite things about it, but as you mentioned earlier, it’s way more stable (clinical?) and less organic sounding than the drift FM analog.

rick_monster · 2017-10-26T16:33:41.973Z

zebra:

heavy oversampling is the typical way in softsynth-land, i think

yeah seems like the crazy 2MHz sampling rate of cyclonix is consistent with the sytrus softsynth (up to 64x oversampling).

So 16x is hardly overkill - might not even cut it… That’s nuts!

anon45924993 · 2017-10-26T16:37:18.813Z

A link for sytrus that @rick_monster mentioned above. Windows only, alas.

image-line.com

Sytrus

Hybrid Synthesizer - subtractive but addictive A disruptive innovation in software synthesis, nothing else compares to the diversity Sytrus packs. FM, RM & Subtractive synthesis - have it all!

A couple softsynths with capable FM implementations that are Mac compatible are NI’s FM8 and Arturia’s Synclavier V.

But I bet there’s some fantastic stuff in the Max and Pd worlds… (well, there certainly is, I just don’t have the links at my fingertips right this minute)

sam · 2017-10-26T16:47:23.454Z

rick_monster:

yeah seems like the crazy 2MHz sampling rate of cyclonix is consistent with the sytrus softsynth (up to 64x oversampling).

I dunno. I have absolutely no backup for this statement, but my mathematical spidey-sense says that each modulator would require a doubling of the sample rate to avoid aliasing. Thus using all 6 operators stacked is 64x oversampling (2^6).

rick_monster · 2017-10-26T16:55:37.725Z

sam:

each modulator would require a doubling of the sample rate to avoid aliasing

definitely true for A.M! cos (w t) * cos (w t) = 1/2 + cos(2w t).

So what’s the expansion of cos(w t + pi cos(w t)) ? time to crack out the fft - here be dragons when it comes to working through the math I think!